You are on page 1of 3

ERRATA FOR

A SHORT COURSE ON SPECTRAL THEORY


William Arveson
Department of Mathematics
University of California
Berkeley CA 94720, USA
arveson@math.berkeley.edu
June, 2003
Abstract. This is a list of corrections I know about. Special thanks go to Launey
Thomas for catching many corrections that I had missed. I welcome additional
feedback from users of this text.
page 5. Exercise (5) is correctly stated, but the hint should be changed as follows.
Hint: Apply Ascolis theorem to KB
1
.
page 11. In Exercise (1), the phrases x
n
X and y X should be replaced with
x
n
E and y E respectively.
page 12. Theorem 1.4.2: the inequality (2) should be
|e|
1
|x| |L
x
| c|x|.
page 13. In Exercise (2), E should be a Banach space.
page 14. Line 3: ...satises |1| = 1.
page 19. Line 7: ...is so, x (x).
page 22. The lower formula displayed in line 12 should be
inf
z
1
,z
2
I
|(x +z
1
)(y +z
2
)| | x|| y|.
page 28. Line 10 should read: K D and z
0
/ D.
page 32. Line 10: that that should be that.
page 32. Line 22, 0 ,=
A
(x) should be 0 /
A
(x).
page 33. In Corollary 1, replace x A with x B.
page 34. Line 20 should read: as the norm |T| of the partition T tends to 0,
where |T| is dened as the maximum of the arc lengths of the segments of C
connecting
k1
to
k
, 1 k n. This denition of |T| implies that |T| must
decrease as the partition T is rened, and simplies Exercise 1 on p. 37.
1
2 WILLIAM ARVESON
page 44. In line 3, p X : [f(p) [ < = 0 should be replaced with
p X : [f(p) [ < = 0.
page 49. Line -5: We do not assume continuity of .
page 51. Line 14: sp(A).
page 56. Exercise (4) should end with the phrase: ...,and which commutes with
the multiplication operator of Exercise (3).
page 56. Exercise (5) should end as follows. Hint: Use the operator U of Exercise
(4), together with the fact that T = X Y where X is the top half and Y is the
bottom half, to show that L
2
(T) can be identied with the direct sum
L
2
(T) L
2
(X, ) L
2
(X, ),
and deduce that operators on this direct sum can be realized as 2 2 matrices over
B(L
2
(X, )). You should write down an explicit unitary operator that makes the
above identication.
page 57. Exercise (7) is inappropriate for this section, since it requires the result
of Theorem 4.1.2 which has not yet been proved. It should probably be moved to
Section 4.1.
Replace What is the commutant of B? with Compare the commutants of A
and B.
page 59. Exercises (4) and (5) should be moved to page 63, as described in the
next item. Reason: they require the fact that von Neumann algebras are generated
by projections, which has not yet been proved.
page 63. Line 12: -representation should be hyphenated.
pp 6364. Three more exercises should be added.
Exercise (3). a). Let X be a compact metric space, let : C(X) B(H) be a
nondegenerate representation of C(X), and let : B(X) B(H) be its extension
to a -representation. Show that for every bounded Borel function f B(X), the
operator (f) belongs to the double commutant (C(X))

of the range of .
b). Deduce that the von Neumann algebra (C(X))

is the norm-closed linear


span of its projections.
Exercise (4). This is Exercise (4) from page 59. Hint: use the the functional
calculus and the preceding exercise.
Exercise (5). Exercise (5) from page 59, verbatim.
page 67. Line 3 of Exercise (1) should read: unitary operator W : H
1
H
2
....
page 72. In line 22, replace see Exeercise (2) below with see Exercise (3) be-
low.
page 78. Line -10: unitilization should be unitalization.
page 91. In line 1, replace Lemma 2.4.3 with Lemma 3.2.7.
page 91. Line 11. Replace is injective with is surjective.
page 98. Line -5: L KB should be L +KB.
page 103. Line -3 should read: commutes with M
f
.
ERRATA 3
page 104. In line -18, replace the measure with the measure .
page 106. In line -12, replace subace with subspace.
page 110. In the Exercises, the notation for Banach limits should be claried as
follows. Line 7 should read: ...we will write (a) or ((a
n
)) for.... The displayed
formula on line 12 should read
(A), ) = ((S
n
AS
n
, ))), , H
2
.
page 110. Interchange f and g in the rst paragraph of the proof of Theorem
4.3.1 as follows. If g C(T) and f L

, ..., we may reduce to the case g =


n
and f L

, n Z.
pages 110111. Line -2: (4.3.1) should be (4.6). Same change in Line 2 of page
111.
page 111. The displayed formula on line 7 needs a change in sign as follows
T
f
n T
f
T

n = S
m
T
f
T
f
S
m
= S
m
T
f
(1 S
m
S
m
).
page 111. In line 20, replace Theorem 4.3.1 with Proposition 4.3.1.
page 116. In the displayed formula on line -4, replace ... = ind(T
fg
+ K) with
... = ind(T
fg
K).
page 117. In line 5, replace Z
+
with Z
+
.
In line -13, replace K =
2
(Z
+
) H with K =
2
(Z
+
) H.
page 118. Line -14 should read ...whose original proof was quite dierent.
page 119. Lines -8 to -7 should read With P
+
B(L
2
) denoting....
page 120. Line 6:

(z) vanishes... should be (z) vanishes....


page 122. In line 11, Lebeggue should be Lebesgue.
page 123. The displayed formula on line 15 should be
[(x)[
2
= [(1

x)[
2
(x

x)(1).
page 124. Line 1 should read ...promotes naturally to a sesquilinear....
page 124. The displayed equation in line 5 should end with
... = x +N = 0 +N.
page 124. Line -6 should begin with (a

) = (a)

, a A.
page 125. In line -4, add a third part to Exercise (2) as follows:
(c) Deduce that any positive element x A (i.e., x is self-adjoint and (x)
[0, )) has a unique positive square root.
page 128. In lines 19 to 19, change sub C

-algebra to C

-subalgebra.

You might also like